introducing the new structural response prediction...
TRANSCRIPT
1
Curt B. Haselton, PhD, PE (and Jack Baker and the SP3 Team)
Professor of Civil Engineering @ CSU, Chico
Co-Founder and CEO @ Seismic Performance Prediction Program (SP3)
www.hbrisk.com
SP3 Software Webinar Series
Introducing the New Structural Response
Prediction Engine
2
© HB Risk Group
All phone lines are muted Questions are highly encourage (answered at end) Handouts are available – presentation slides and one-page
summary of the new method Webinar is recorded and video will be distributed Please take a brief survey before signing off at end of webinar
Housekeeping
3
© HB Risk Group
Questions:• Please use questions tab
and we will address as many as we can at the end.
• For further questions, contact Dr. Haselton directly ([email protected]) or contact the SP3 team ([email protected]).
Housekeeping
4
© HB Risk Group
This is the first webinar in our series, so please watch for announcements for our next three webinars!
Webinar Series:1) The new SP3 Structural Response Prediction Engine2) Risk assessment of wood light-frame buildings using SP33) Risk assessment of tilt-up buildings using SP34) SP3: A consistent platform for building-specific seismic risk
assessment for low-levels to high-levels of building input information
Housekeeping
5
Curt B. Haselton, PhD, PE (and Jack Baker and the SP3 Team)
Professor of Civil Engineering @ CSU, Chico
Co-Founder and CEO @ Seismic Performance Prediction Program (SP3)
www.hbrisk.com
SP3 Software Webinar Series
Introducing the New Structural Response
Prediction Engine
6
© HB Risk Group
Brief overview of FEMA P-58 and the SP3 software The need for rapid and advanced seismic risk evaluation The new Structural Response Prediction Engine:
• Overview of approach• Detail of what is under the hood• What the Engine now enables
Current and next steps for SP3 Research & Development Questions
Outline for Today
7
© HB Risk Group
FEMA P-58 is a probabilistic performance prediction methodology (15 years of effort to date, ~$16M invested by FEMA)
FEMA P-58 is tailored for building-specific analysis (in contrast to most risk assessment methods that give results by building class)
FEMA P-58 output results:• Repair costs• Repair time• Safety: Fatalities & injuries
FEMA P-58 Overview
8
© HB Risk Group
FEMA P-58 Overview
Ground Motion Hazard
Component DamageEconomic Loss
Casualties
Repair Time
Structural Responses
SP3 uses FEMA P-58 and adds much more:
9
© HB Risk Group
FEMA P-58 provides the comprehensive and standardized
building-specific risk assessment (with ~$16M to develop).
SP3 provides a user-friendly software to integrate all steps in a FEMA P-58
risk assessment.
The initial assessment should take a couple hours and not days or
weeks.
SP3 Software Compliments FEMA P-58
10
© HB Risk Group
Our Goal: To enable and facilitate rapid and advanced building-specific seismic risk assessments for all buildings(i.e. building-specific vulnerability curves).
Expected Outcomes: We believe that this will (a) facilitate the design of moreresilient buildings and (b) enable better risk-relateddecision making (e.g. insurance risk, mortgage risk, etc.).
Our Goal with SP3
11
© HB Risk Group
Current uses of SP3:• Licensed by ~80+% of the large west-coast structural engineering firms, and
has been expanding from there.
• Currently being used for:
• Resilient design of new buildings (municipal buildings, court houses, etc.)
• Retrofit of existing buildings
• Assessments of special facilities (EOCs, manufacturing, museums, etc.)
• Mortgage risk assessments [needs rapid risk methods]
• Investment risk assessments [needs rapid risk methods]
• Insurance risk assessments [needs rapid risk methods]
Current Uses of SP3 Software
12
© HB Risk Group
Our Goal: To enable and facilitate rapid and advanced building-specific seismic risk assessments for all buildings(i.e. building-specific vulnerability curves).
Expected Outcomes: We believe that this will (a) facilitate the design of moreresilient buildings and (b) enable better risk-relateddecision making (e.g. insurance risk, mortgage risk, etc.).
Our Goal with SP3
13
© HB Risk Group
Our Goal: To enable and facilitate rapid and advanced building-specific seismic risk assessments for all buildings(i.e. building-specific vulnerability curves).
Expected Outcomes: We believe that this will (a) facilitate the design of moreresilient buildings and (b) enable better risk-relateddecision making (e.g. insurance risk, mortgage risk, etc.).
Our Goal with SP3
14
© HB Risk Group
Ground Motion Hazard
Component DamageEconomic Loss
Casualties
Repair Time
Structural Responses
The Need for Rapid Risk Evaluations
15
© HB Risk Group
Structural Responses
SP3 Structural Response Prediction ENGINE
“We do the nonlinear dynamic structural analysis for you.”
The Need for Rapid Risk Evaluations
[and contrast to FEMA P-58 “Simplified Method”]
16
© HB Risk Group
Brief overview of FEMA P-58 and the SP3 software The need for rapid and advanced seismic risk evaluation The new Structural Response Prediction Engine:
• Overview of approach• Detail of what is under the hood• What the Engine now enables
Current and next steps for SP3 Research & Development Questions
Outline for Today
17
© HB Risk Group
Brief overview of FEMA P-58 and the SP3 software The need for rapid and advanced seismic risk evaluation The new Structural Response Prediction Engine:
• Overview of approach• Detail of what is under the hood• What the Engine now enables
Current and next steps for SP3 Research & Development Questions
Outline for Today
18
© HB Risk Group
We first assembled a large set of structural models and simulated nonlinear responses for many ground motion levels.
Approach to the SP3 Str. Response Prediction Engine
19
© HB Risk Group
We first assembled a large set of structural models and simulated nonlinear responses for many ground motion levels.
Approach to the SP3 Str. Response Prediction Engine
20
© HB Risk Group
We first assembled a large set of structural models and simulated nonlinear responses for many ground motion levels.
Approach to the SP3 Str. Response Prediction Engine
0.2
12
0.3
0.4
10
0.5
12
0.6
IDR
Cor
rela
tion
8
0.7
10
0.8
Story Number
6 8
0.9
Story Number
6
1
44
22
0 0
21
© HB Risk Group
We first assembled a large set of structural models and simulated nonlinear responses for many ground motion levels.
Approach to the SP3 Str. Response Prediction Engine
22
© HB Risk Group
Engineering Demand Parameters for ~100 ground motions
(drifts and floor accelerations)
SP3 Structural Response Prediction ENGINE“We do the nonlinear dynamic structural analysis for you.”
Approach to the SP3 Str. Response Prediction Engine
23
© HB Risk Group
Structural Responses
SP3 Structural Response Prediction ENGINE
“We do the nonlinear dynamic structural analysis for you.”
SP3 Auto-Population ENGINE
Structural Inputs:- Building strength
- T1, T2, and T3
- First three mode shapes
Inputs:- Building system and height- Building age (and/or code)
- Building location- Other optional site/design info.
Database: Structural responses from hundreds of nonlinear structural models with millions of nonlinear response-history analyses.
Approach to the SP3 Str. Response Prediction Engine
24
© HB Risk Group
Structural Responses
Inputs for New “Low Data” Option:- Building system and height- Building age (and/or code)
- Building location- Other optional site/design info.
Approach to the SP3 Str. Response Prediction Engine
Previous “Full Data” Option:
- Create a structural model, run nonlinear response-history analyses, and input results
manually.
Inputs for New “Intermediate Data” Option:
- Building strength- T1, T2, and T3
- First three mode shapes
25
© HB Risk Group
Approach to the SP3 Str. Response Prediction Engine
26
© HB Risk Group
This was an overview of how we created the new SP3 Response Engine. Let’s now get into more of the detail.
Creating the SP3 Str. Response Prediction Engine
27
© HB Risk Group
Example data set for a 12-story Reinforced Concrete Special Moment Frame in high-seismic California.
Creating the SP3 Str. Response Prediction Engine
28
© HB Risk Group
Example data set for a 12-story Reinforced Concrete Special Moment Frame in high-seismic California.
Creating the SP3 Str. Response Prediction Engine
29
© HB Risk Group
Example data set for a 12-story Reinforced Concrete Special Moment Frame in high-seismic California.
Creating the SP3 Str. Response Prediction Engine
30
© HB Risk Group
Creating the drift prediction algorithm:
Step 1: Use nonlinear response data set to get correct building global displacement demands.
• Start with ASCE 41 target displacement.• Correct this based on the database of structural responses.
Creating the SP3 Str. Response Prediction Engine
𝛿𝛿𝑡𝑡 = 𝐶𝐶0𝐶𝐶1 𝐶𝐶2𝑆𝑆𝑎𝑎𝑇𝑇𝑒𝑒2
4𝜋𝜋2 𝑔𝑔
0 1 2 3 4 5 6
S (Sa(Teff)/Vmax)
0
0.005
0.01
0.015
RD
R
Observered RDR 1015
Predicted RDR 1015
ASCE41 RDR 1015
31
© HB Risk Group
Creating the drift prediction algorithm:
Step 2: Since peak story drifts over height do not occur all at the same time, use the database data to get the average story drifts correct.
Creating the SP3 Str. Response Prediction Engine
0 1 2 3 4 5 6
S (Sa(T1)/Vmax
0
0.005
0.01
0.015
Ave
rage
IDR
Observed ave IDR: model 1015
Predicted ave IDR: model 1015
32
© HB Risk Group
0 0.005 0.01 0.015 0.02 0.025 0.03
IDR
0
2
4
6
8
10
12
Sto
ry
Opensees IDR: S = 0.80Elastic IDR: S = 0.80
Opensees IDR: S = 1.82Elastic IDR: S = 1.82
Opensees IDR: S = 2.84Elastic IDR: S = 2.84
Opensees IDR: S = 3.86Elastic IDR: S = 3.86
Opensees IDR: S = 4.89Elastic IDR: S = 4.89
Creating the drift prediction algorithm:
Step 3: Do a three-mode modal analysis to compute elastic story drift profiles. This is based on T1-T3 and φ1-φ3 either from an input or auto-populated based on building type and design information.
Creating the SP3 Str. Response Prediction Engine
33
© HB Risk Group
Creating the drift prediction algorithm:
Step 4: Use the response database to modify the response to account for inelastic effects.
• This includes an adjustment to the global drifts, to account for inelastic damping effects (when they occur).
• This also handles where the drifts localize over height (which depend on structural system and design approach, e.g. strong column-weak beam).
Creating the SP3 Str. Response Prediction Engine
34
© HB Risk Group
0 0.005 0.01 0.015 0.02 0.025 0.03
IDR
0
2
4
6
8
10
12
Sto
ry
Opensees IDR: S = 0.80Elastic IDR: S = 0.80
Predicted IDR: S = 0.80
Opensees IDR: S = 1.82Elastic IDR: S = 1.82
Predicted IDR: S = 1.82
Opensees IDR: S = 2.84Elastic IDR: S = 2.84
Predicted IDR: S = 2.84
Opensees IDR: S = 3.86Elastic IDR: S = 3.86
Predicted IDR: S = 3.86
Opensees IDR: S = 4.89Elastic IDR: S = 4.89
Predicted IDR: S = 4.89
Creating the drift prediction algorithm:
Step 4: Use the response database to modify the response to account for inelastic effects.
Creating the SP3 Str. Response Prediction Engine
35
© HB Risk Group
The drift prediction algorithm is now done! Now, let’s look at creating the floor acceleration prediction
algorithm.
Creating the SP3 Str. Response Prediction Engine
36
© HB Risk Group
0 0.5 1 1.5
Acceleration (g)
0
2
4
6
8
10
12
Floo
r
Opensees PFA: S = 0.80
Opensees PFA: S = 1.82Elastic PFA: S = 1.82
Opensees PFA: S = 2.84Elastic PFA: S = 2.84
Opensees PFA: S = 3.86Elastic PFA: S = 3.86
Opensees PFA: S = 4.89Elastic PFA: S = 4.89
Creating the floor acceleration prediction algorithm:
Step 1: Do a three-mode modal analysis to estimate elastic floor accelerations.
Creating the SP3 Str. Response Prediction Engine
𝑃𝑃𝑃𝑃𝑃𝑃𝑛𝑛 = �𝑖𝑖=1
3
Γ𝑖𝑖𝜙𝜙𝑖𝑖,𝑛𝑛�̈�𝑢(𝑡𝑡)𝑖𝑖
37
© HB Risk Group
0 0.5 1 1.5
Acceleration (g)
0
2
4
6
8
10
12
Floo
r
Opensees PFA: S = 0.80Predicted PFA: S = 0.80
Elastic PFA: S = 0.80
Opensees PFA: S = 1.82Predicted PFA: S = 1.82
Elastic PFA: S = 1.82
Opensees PFA: S = 2.84Predicted PFA: S = 2.84
Elastic PFA: S = 2.84
Opensees PFA: S = 3.86Predicted PFA: S = 3.86
Elastic PFA: S = 3.86
Opensees PFA: S = 4.89Predicted PFA: S = 4.89
Elastic PFA: S = 4.89
Creating the floor acceleration prediction algorithm:
Step 2: Use structural response database to capture the effects of inelastic behavior.
Creating the SP3 Str. Response Prediction Engine
38
© HB Risk Group
Response correlations:• Getting these right is crucial to getting a meaningful estimate of
variability in the final risk prediction (e.g. getting full curve to estimate things like 90th percentile of losses).
Creating the SP3 Str. Response Prediction Engine
0.2
12
0.3
0.4
10
0.5
12
0.6
IDR
Cor
rela
tion
8
0.7
10
0.8
Story Number
6 8
0.9
Story Number
6
1
44
22
0 0
39
© HB Risk Group
To summarize…
Summary of SP3 Structural Response Engine
40
© HB Risk Group
Summary of SP3 Structural Response Engine
This new ENGINE enables a high-fidelity SP3 risk assessment without
needing to build a nonlinear structural model.
This does not replace modeling by a structural engineer and
some applications of SP3 will still include modeling, in order to reduce the uncertainty in the risk
assessment (e.g. for resilient design of new buildings).
However, this opens the door to high-fidelity risk assessments for the
many cases where a structural model is not feasible (e.g. initial structural design, insurance risk,
mortgage risk, investment risk, etc.).
41
© HB Risk Group
Current scoping limitations of SP3 Structural Response Prediction Engine:
• Applies up to 25 stories (wood up to 5-story with up to 2-story pedestal)• Calibrated for both ductile and non-ductile buildings
(e.g. RC SMF, OMF, and 1960’s RC frames)• Regular buildings (and currently being extended)• Data set supports low levels of ground motion (elastic) to high levels
(highly nonlinear); include response up to 25% collapse rate (so no meaningful limits on ductility level)
• Response correlations are carefully tracked, so a full distribution of risk results can be reliably provided (e.g. 90% percentile).
• Most aspects of the Engine are generic to all structural systems (e.g. modal analysis), but inelastic factors are currently based on frame data (and are currently being extended).
Summary of SP3 Structural Response Engine
42
© HB Risk Group
With substantial funding from the National Science Foundation ($980k to date), we are also continuing further development for rapid and advanced building-specific risk assessment.
The research focuses are:
Further extending and refining the Structural Response Prediction Engine.
Rapid assessment of wood light-frame buildings (done).
Rapid assessment of tilt-up buildings.
Next Research and Development Steps
43
© HB Risk Group
Next steps on the SP3 Response Engine:• Extend to taller – up to 40 stories• Refine inelastic factors for additional structural systems – e.g.
current in-progress project looking at buckling-restrained braced frame buildings up to 40-stories
• Extend to irregular – vertical weak story and plan torsion (structural analyses are already completed in the database)
Next Research and Development Steps
44
© HB Risk Group
Wood light frame (covers single- and multi-family, up to 5-stories of wood and 1-2 stories of podium) [method is complete]
Next Research and Development Steps
45
© HB Risk Group
Tilt-up buildings (rigid-wall flexible diaphragm) – inclusive of all building era’s, focus on tilt-up panels with wood diaphragms (i.e. buildings is western half of the U.S.)
Next Research and Development Steps
46
© HB Risk Group
Closing and Questions
Thank you for your time. Our goal is to support adoption of resilience-based design and risk
assessment, and we welcome feedback and suggestions.
Time for questions!
Curt Haselton: [email protected], Direct: (530) 514-8980
Jack Baker: [email protected]
www.hbrisk.com
47
© HB Risk Group
Please watch for announcements for our next three webinars!
Webinar Series:1) The new SP3 Structural Response Prediction Engine2) Risk assessment of wood light-frame buildings using SP33) Risk assessment of tilt-up buildings using SP34) SP3: A consistent platform for building-specific seismic risk
assessment for any levels of building information
Upcoming Webinars in this Series
48
© HB Risk Group
Questions:• Please use questions tab
and we will address as many as we can at the end.
• For further questions, contact Dr. Haselton directly ([email protected]) or contact the SP3 team ([email protected]).
Closing and Questions